TOF/T0 Update
Analysis Meeting 3/4/2003
Progress on the creation of TOF PID plots is slowed down by problems associated with the determination of the collision time.
The accuracy of the collision time can be determined by looking at the resolution of the reconstructed vertex.
We can determine the collision time in two different ways:
T0 OR information:
Works but is limited in resolution due the TDC resolution of 100 ps/ch and the accuracy of the relative alignment in time of the T0s.
Based on the vertex resolution, the collision time resolution derived from the ORs is 160 ps (1 s).
Use the individual T0 time information:
If proper delay corrections are used, the T0 information should be able to exceed the resolution of the ORs.
The main problem is determining the proper delay corrections.

TOF/T0 Update
Analysis Meeting 3/4/2003
Since the last bi-weekly meeting we have completed the following tasks:
Studied drifts in TDC gains and determined that this is a small effect.  The changes in gain can not be responsible for the excess drifts we have observed.
Implemented a new calibration procedures that corrects for drifts associated with for example the crossing clock.  This procedures looks at the average time of prompt peak of T0N and the T0P counters and calculates a common shift required to align the centroids of the T0N and the T0P counters with the expected position of the prompt peak.  The same procedure will also work for the TOF since it does not require high statistic in each individual channel.
Determined first order slewing parameters for the T0N sensors using beam-gas interactions.  Work on the T0P sensors is in progress.
Continued DST production required for TOF studies.

TOF/T0 Update
New time-delay calibrations
The new calibration procedure forces the mean of the time distribution of T0N and T0P to be located at the required time.
The time peaks are rather wide (1.2 ns) and the centroid of these distributions must be determined with an accuracy of better than 100 ps.
The distributions contain both primary and secondary contributions and we assume the relative strength of these contributions is the same for all counters on one array.

T0 Vertex Studies:
Comparing Spec Vertex with Oct dE Vertex
Our studies of T0 vertexing were carried out using run 10920.
Run 10920 is a long run, with about 63 sequences, taken on 1/28/03 with a B+ field.
We have used Spec Vertex to compare the silicon vertex with the T0 vertex. and selected only events for which the Spec Vertex is located between ±20 cm.
For events with a valid spec vertex the Oct dE vertex and Spec Vertex agree within a resolution of 1.0 cm (1 s).

T0 Vertex Studies:
Comparing Spec Vertex with T0 OR Vertex
The T0 OR Vertex is obtained from the OR information obtained using TDC 8:
T0 OR Vertex = 29.979*((tExtra[19]  -  311.87) - (tExtra[20] - 302.65))/2.0
The vertex resolution is 4.8 cm (1 s).
This implies a resolution of the T0 Ors of 225 ps (1 s).  This resolution is most likely dominated by the alignment accuracy.

T0 Vertex Studies:
T0N0 - T0P0 Vertex vs Spec Vertex
The vertex resolution based on T0N0 and T0P0 is 7.0 cm (1 s).
This vertex resolution would imply a T0 resolution of 330 ps (1 s).  This resolution must be dominated by secondaries.

T0 Vertex Studies:
T0N5 - T0P0 Vertex vs Spec Vertex
The vertex resolution based on T0N5 and T0P0 is 15.1 cm (1 s).
This vertex resolution would imply a T0 resolution of 710 ps (1 s).  This resolution must be dominated by secondaries.

T0 Vertex Studies:
T0N5 - T0P5 Vertex vs Spec Vertex
The vertex resolution based on T0N0 and T0P0 is 7.4 cm (1 s).
This vertex resolution would imply a T0 resolution of 350 ps (1 s).  This resolution must be dominated by secondaries.

T0 Vertex Studies:
T0N0 - T0P5 Vertex vs Spec Vertex
The vertex resolution based on T0N5 and T0P0 is 19.2 cm (1 s).
This vertex resolution would imply a T0 resolution of 910 ps (1 s).  This resolution must be dominated by secondaries.

T0 Vertex Studies:
T0N5/T0P5 Vertex vs Spec Vertex
T0 Vertex Studies:
T0N2/T0P2 Vertex vs Spec Vertex
T0 Vertex Studies:
T0N5 - T0P5 Vertex vs Spec Vertex
Although there is small excess of yield at dz = 0 cm, it is clear that we see only a small number of primary-primary coincidences.
If the spectra are dominated by secondaries, then there is no reason to assume that the time distributions for the individual T0 counters should be centered at the same time.

T0 Vertex Studies
How can we explain this?
Bug in software when T0 event is created?
Not very likely since it was written by a Dutchman, but investigated anyway by looking at the raw FB data.  Correlations are also clearly visible in these data.
Effect of magnetic field?
No; runs with no magnetic field show the same effect.
Conclusion: this is a real effect.

T0 Vertex Studies
If all T0 counters are properly aligned in time,  we expect that by using the time of the first T0 to fire we should be able to match or exceed the performance of the T0 OR vertex.
The results obtained with these first T0 signals are very poor, clearly showing that the T0 counters are not properly aligned in time.

T0 Vertex Studies
What can cause this asymmetry?
It is not the magnetic field.
It is not the floor.
It is not the software.
It is not the calculated time delay (this effect the position of the vertex distribution for a specific counter combination, not the width).
At the moment, we have no clear indication what effect is responsible for the observed asymmetry.  Monte-Carlo studied are in progress to see if this effect is also present in the simulated data.

T0 Vertex Studies:
Primary particles can not produce asymmetry.
T0 Vertex Studies: Off-axis particle production can produce (small) asymmetry.
T0 Vertex Studies
Summary
Our alignment procedure for the T0 counters clearly does not work.
It appears that our T0 time distributions are dominated by secondaries, and that the actual secondaries effect different counters in different ways.
We will first need to verify whether or not the effect we observe in the data is also present in the Monte-Carlo data.